Abstract
Free full text
Evidence of cellulose metabolism by the giant panda gut microbiome
Associated Data
The giant panda genome codes for all necessary enzymes associated with a carnivorous digestive system but lacks genes for enzymes needed to digest cellulose, the principal component of their bamboo diet. It has been posited that this iconic species must therefore possess microbial symbionts capable of metabolizing cellulose, but these symbionts have remained undetected. Here we examined 5,522 prokaryotic ribosomal RNA gene sequences in wild and captive giant panda fecal samples. We found lower species richness of the panda microbiome than of mammalian microbiomes for herbivores and nonherbivorous carnivores. We detected 13 operational taxonomic units closely related to Clostridium groups I and XIVa, both of which contain taxa known to digest cellulose. Seven of these 13 operational taxonomic units were unique to pandas compared with other mammals. Metagenomic analysis using ~37-Mbp contig sequences from gut microbes recovered putative genes coding two cellulose-digesting enzymes and one hemicellulose-digesting enzyme, cellulase, β-glucosidase, and xylan 1,4-β-xylosidase, in Clostridium group I. Comparing glycoside hydrolase profiles of pandas with those of herbivores and omnivores, we found a moderate abundance of oligosaccharide-degrading enzymes for pandas (36%), close to that for humans (37%), and the lowest abundance of cellulases and endohemicellulases (2%), which may reflect low digestibility of cellulose and hemicellulose in the panda's unique bamboo diet. The presence of putative cellulose-digesting microbes, in combination with adaptations related to feeding, physiology, and morphology, show that giant pandas have evolved a number of traits to overcome the anatomical and physiological challenge of digesting a diet high in fibrous matter.
Access to dietary resources shapes animal evolution (1). Early on, animals lost the ability to synthesize many key compounds, and instead this function is performed by symbionts (2). For example, microbial symbionts assist with extracting nutrients from food and key compounds from the environment, and also synthesize necessary metabolic compounds (1). Gut microbiota share specialized relationships with their hosts, and advances in genomics are revealing the dynamics of these relationships (3). Recent developments in culture-independent methodologies based on large-scale comparative analyses of microbial small-subunit ribosomal RNA genes (16S ribosomal RNA) and metagenomics have revealed the extent of microbial diversity and metabolic potential in greater detail (2–7). These techniques can now be applied to animals that have acquired a profoundly new diet, presenting an opportunity to investigate host physiological and microbial systems in an evolutionary context.
The giant panda (Ailuropoda melanoleuca) is well known for dietary oddities: a bamboo specialist within the mammalian order Carnivora possessing a gastrointestinal tract typical of carnivores. It consumes ~12.5 kg of this highly fibrous plant each day (8), but because it lacks the long intestinal tract characteristic of other herbivores, extensive fermentation is not possible (9). Giant pandas digest only ~17% of dry matter consumed (8), and have low digestion coefficients for bamboo hemicelluloses (27%) and celluloses (8%) (9). Indeed, the giant panda genome codes for all necessary enzymes associated with a carnivorous digestive system, but lacks the enzyme homologs needed for cellulose digestion (10). Although the giant panda can use noncellulosic material from the bamboo diet using enzymes coded in its own genome, digestion of cellulose and hemicellulose is impossible based on the panda's genetic composition, and must be dependent on gut microbiome. However, previous research using culture methods and small-scale sequencing identified three predominant bacteria from the panda gut—Escherichia coli, Streptococcus, and Enterobacteriaceae—none of which aids in cellulose digestion (11–13). Thus, an incomplete understanding of the gut microbial ecosystem in this interesting and high-profile species remains because of restrictions in methodology and past reliance on studies of captive animals.
We undertook a large-scale analysis of 16S rRNA gene sequences to profile microbial flora inhabiting the digestive system of giant pandas and used a metagenomic approach based on next-generation de novo sequencing to identify functional attributes encoded in the gut microbiome. A total of 5,636 near–full-length 16S rRNA gene segments were amplified from fecal samples of seven wild and eight captive giant pandas. After exclusion of 74 putative chimeric and 30 chloroplast sequences, 5,522 sequences were retained for analysis. Using a minimum identity of 97% as the threshold for any sequence pair, we identified 85 bacterial operational taxonomic units (OTUs), 14 of which were previously undescribed (Fig. 1A and SI Appendix, Table S1). Coverage of 99% was obtained across existing bacterial clone libraries, and we are confident that our dataset presents the most comprehensive assessment of gut microbes in this species based on the rarefaction method in DOTUR (14) (SI Appendix, Fig. S1A). The majority of microbes were members of the Firmicutes (62 OTUs, 4,633 sequences, 83.8% of the total of 5,522 sequences) and Proteobacteria (12 OTUs, 871 sequences, 15.8% of the total sequences), with the remainder belonging to the phyla Actinobacteria, Bacteroidetes, Cyanobacteria, and Acidobacteria (Fig. 1 and SI Appendix, Fig. S2 and Table S1). Within the Firmicutes, 33 OTUs (60.8% of the total of 5,522 sequences) were members of the class Clostridia and 29 OTUs (23.0% of total sequences) belonged to the class Bacilli. The high proportion of Firmicutes OTUs found in the gut of giant pandas differs from the findings in previous studies attempting to characterize giant panda gut microbes (11–13) and is notably similar to that found in herbivorous mammals (2, 6). However, the species richness in the giant panda was lower than that seen in previous work on mammalian herbivores and nonherbivorous carnivores (SI Appendix, Fig. S1B), perhaps due to the panda's unique bamboo diet and simple digestive system.
Microbial flora of giant pandas. (A) Neighbor-joining tree containing one representative of each of 85 OTUs. Colors represent different phyla: yellow, Clostridia class of Firmicutes; blue, Bacilli class of Firmicutes; purple, Proteobacteria; gray, Actinobacteria, Bacteroidetes, Cyanobacteria, and Acidobacteria. seqs, sequences. All bootstrap values >80% are shown in this tree. (B) Relative abundance of sequences from wild and captive giant panda fecal samples; other Firmicutes belong to the Bacilli class. (C) Percentage of sequences from each fecal sample assigned to different phyla. W1–W7, wild giant panda samples; C1–C8, captive giant panda samples. Numbers below the sample number are total sequences from each individual.
The proportion of 16S rRNA gene sequences belonging to the class Clostridia of Firmicutes were high for both wild (73% of the total of 3361 sequences) and captive (42% of the total of 2161 sequences) giant pandas (Fig. 1 B and C). In Clostridia, we detected 13 OTUs closely related to Clostridium group I (10 OTUs, 1,457 sequences, 26.4% of the total sequences) and XIVa (3 OTUs, 185 sequences, 3.3% of the total sequences) (SI Appendix, Fig. S3), groups that contain organisms known to digest cellulose (15–18). Moreover, we found that 7 of these 13 OTUs were unique to giant pandas compared with other mammals, including Artiodactyla, Perissodactyla, Rodentia, Diprotodontia, and primates (2, 19–21) (SI Appendix, Table S2).
Using gene functional classification in a metagenomic analysis of fecal samples from three wild individuals (W1, W2, and W5), we found sequences homologous to genes coding cellulase, β-glucosidase, xylan 1,4-β-xylosidase, and endo-1,4-β-xylanase. The de novo assembly (22) for Illumina Genome Analyzer sequence data resulted in a set of ~37 Mbp (SI Appendix, Table S4). We used MetaGene (23) to predict ORFs in our contigs and found 49,844 genes occupying nearly 32 Mbp. Of these genes, 44% possessed a complete ORF (SI Appendix, Tables S5 and S6). We determined the taxonomic level of the gene catalog with their protein sequences by the lowest common ancestor (LCA)-based algorithm implemented in MEGAN (24). The results again showed that the majority of microbes (71%) belong to the phylum Firmicutes (Fig. 2A). We also compared the catalog with the Kyoto Encyclopedia of Genes and Genomes (KEGG) database and the Clusters of Orthologous Groups (COG) databases to assess the functional capacities present in bacterial metagenome (Fig. 2 B–D and SI Appendix, Figs. S4 and S5). Our analysis revealed that half of these predicted genes coding for cellulose and hemicellulose-digesting enzymes are found in species within the Clostridium genus (SI Appendix, Table S7). For example, 2 of 12 putative cellulase genes in this study are homologous (E-value: 2.00E-115 and 6.00E-29, respectively) to genes from Clostridium butyricum of Clostridium group I (SI Appendix, Table S7), and 1,968 of 49,844 genes were assigned to this species. Of interest, OTU54 and OTU55 have a close phylogenetic relationship with C. butyricum (SI Appendix, Fig S3), and OTU54 was found in eight individuals [W1 (4.1%), W2 (13.4%), W3 (0.3%), W4 (1.5%), W5 (43.4%), W6 (3.5%), C3 (12.1%), and C8 (1.2%)] (SI Appendix, Table S1). We also found that 21 putative genes are homologous to genes coding xylan 1,4-β-xylosidase from Clostridium (SI Appendix, Table S7). KEGG metabolic pathways provide a highly integrated picture of global gut cell metabolism (Fig. 2B), suggesting that the gut microbiome of giant pandas has enriched activity for metabolism of carbohydrates, amino acids, xenobiotics, nucleotides, glycans, vitamins, and lipids (SI Appendix, Fig. S5). Moreover, comparative COG functional analysis of our data with several gastrointestinal or rumen metagenomes (6, 7, 20, 25) demonstrated metabolic similarities among panda, wallaby, cow, termite, and human, associated mainly with carbohydrates, amino acids, and DNA metabolism (SI Appendix, Fig. S6).
The function of gut metagenomes from three wild giant pandas (W1, W2, and W5). (A) Gene taxonomic assignment. NA, no matched data or no taxon information for matched genes. “Others” includes bacteria in other classes, such as Bacteroidetes and Actinobacteria. (B) Projection of the gut genome on the KEGG pathways using iPath (26). (C) Pathway of cellulose digestion. The box indicates that 1,4-β-cellobiosidase was partially found in the unassembly short-read sequences set. (D) Functional composition of the gut metagenome in the COG database.
Database searches for glycoside hydrolase (GH) genes were performed using HMMER HMM-based sequences search with Pfam hidden Markov models (HMMs), Pfam v25.0 and HMMER v3.0. These analyses found 448 genes from 44 different GH families in the Carbohydrate-Active Enzymes (CAZy) database (Table 1). Comparing GH profiles with those of other herbivores (wallaby, termite, cow) and an omnivore (human) (6, 7, 20, 25), we found that the abundance of oligosaccharide-degrading enzymes (36%) from giant pandas was close to that in human (37%) and lower than that in cow (57%); however, the abundance of cellulases and endohemicellulases was the lowest of the herbivores (SI Appendix, Table S8). These findings may reflect the low digestibility of the cellulose and hemicellulose present in the unique bamboo diet of giant pandas.
Table 1.
Inventory of putative GHs identified in the giant panda gut microbiome
| CAZy family | Pfam HMM name | Pfam accession | Known activities | Giant panda gut community |
| GH1 | Glyco_hydro_1 | PF00232 | β-Glucosidase, β-galactosidase, β-mannosidase, others | 101 |
| GH2 | Glyco_hydro_2_C | PF00703 | β-Galactosidase, β-mannosidase, others | 1 |
| GH3 | Glyco_hydro_3 | PF00933 | β-1,4-Glucosidase, β-1,4-xylosidase, β-1,3-glucosidase, α-larabinofuranosidase, others | 18 |
| GH4 | Glyco_hydro_4 | PF02056 | α-Glucosidase, a-galactosidase, α-glucuronidase, others | 6 |
| GH5 | Cellulase | PF00150 | Cellulase, β-1,4-endoglucanase, β-1,3-glucosidase, β-1,4-endoxylanase, β-1,4-endomannanase, others | 3 |
| GH8 | Glyco_hydro_8 | PF01270 | Cellulase, β-1,3-glucosidase, β-1,4-endoxylanase, β-1,4-endomannanase, others | 2 |
| GH10 | Glyco_hydro_10 | PF00331 | Xylanase, β-1,3-endoxylanase | 2 |
| GH13 | Alpha-amylase | PF00128 | α-Amylase, catalytic domain, and related enzymes | 35 |
| GH16 | Glyco_hydro_16 | PF00722 | β-1,3(4)-Endoglucanase, others | 9 |
| GH18 | Glyco_hydro_18 | PF00704 | Chitinase, endo-β-N-acetylglucosaminidase, noncatalytic proteins | 1 |
| GH20 | Glyco_hydro_20 | PF00728 | β-Hexosaminidase, lacto-N-biosidase | 7 |
| GH23 | SLT | PF01464 | G-type lysozyme, peptidoglycan lytic transglycosylase | 20 |
| GH24 | Phage_lysozyme | PF00959 | Lysozyme | 1 |
| GH25 | Glyco_hydro_25 | PF01183 | Lysozyme | 9 |
| GH27 | Melibiase | PF02065 | α-Galactosidase, α-N-acetylgalactosaminidase, isomalto-dextranase | 36 |
| GH28 | Glyco_hydro_28 | PF00295 | Polygalacturonase, rhamnogalacturonase, others | 1 |
| GH29 | Alpha_L_fucos | PF01120 | α-l-fucosidase | 1 |
| GH31 | Glyco_hydro_31 | PF01055 | α-Glucosidase, α-xylosidase, others | 9 |
| GH32 | Glyco_hydro_32N | PF00251 | Levanase, invertase, others | 13 |
| GH35 | Glyco_hydro_35 | PF01301 | β-Galactosidase | 4 |
| GH36 | No Pfam | No Pfam | α-Galactosidase, α-N-acetylgalactosaminidase | 33 |
| GH37 | Trehaalse | PF01204 | α,α-trehalase | 2 |
| GH38 | Glyco_hydro_38 | PF01074 | α-Mannosidase | 10 |
| GH39 | Glyco_hydro_39 | PF01229 | β-Xylosidase, α-l-iduronidase | 9 |
| GH42 | Glyco_hydro_42 | PF02449 | β-Galactosidase | 18 |
| GH51 | No Pfam | No Pfam | Endoglucanase, α-l-arabinofuranosidase | 5 |
| GH55 | No Pfam | No Pfam | Exo-1,3-glucanase, endo-1,3-glucanase | 1 |
| GH58 | No Pfam | No Pfam | Endo-N-acetylneuraminidase or endo-sialidase | 1 |
| GH65 | Glyco_hydro_65m | PF03632 | Trehalase, maltose phosphorylase, trehalose phosphorylase | 1 |
| GH67 | Glyco_hydro_67M | PF03648 | α-Glucuronidase, others | 2 |
| GH70 | Glyco_hydro_70 | PF02324 | Dextransucrase, alternansucrase | 9 |
| GH72 | Glyco_hydro_72 | PF03198 | β-1,3-glucanosyltransglycosylase | 3 |
| GH73 | Glucosaminidase | PF01832 | Peptidoglycan hydrolase with endo-β-N-acetylglucosaminidase specificity | 28 |
| GH77 | Glyco_hydro_77 | PF02446 | 4-α-Glucanotransferase, amylomaltase | 12 |
| GH78 | Bac_rhamnosid | PF05592 | α-l-Rhamnosidase | 2 |
| GH85 | Glyco_hydro_85 | PF03644 | Endo-β-N-acetylglucosaminidase | 4 |
| GH88 | Glyco_hydro_88 | PF07470 | d-4,5 Unsaturated β-glucuronyl hydrolase | 1 |
| GH89 | NAGLU | PF05089 | Alpha-N-acetylglucosaminidase | 2 |
| GH90 | No Pfam | No Pfam | Endorhamnosidase | 1 |
| GH94 | No Pfam | No Pfam | Cellobiose phosphorylase, chitobiose phosphorylase, cellodextrin phosphorylase | 7 |
| GH95 | No Pfam | No Pfam | α-l-Fucosidase | 4 |
| GH101 | No Pfam | No Pfam | Endo-α-N-acetylgalactosaminidase | 6 |
| GH104 | Phage_lysozyme | PF00959 | Peptidoglycan lytic transglycosylase | 1 |
| GH109 | No Pfam | No Pfam | α-N-acetylgalactosaminidase | 7 |
CAZy family and known activities are from the Carbohydrate Active Enzymes database (http://www.cazy.org/) (32). Pfam HMM-based sequences name and Pfam accession are from the Pfam database (http://pfam.sanger.ac.uk/) (33), a large collection of protein families, each represented by multiple sequence alignments and hidden Markov models. Giant panda gut community: Number of predicted genes match the Pfam domains or “non-Pfam” representative sequences.
Our metagenomic analysis and 16S rRNA gene data confirm the presence of putative cellulose-metabolizing symbionts in this little-studied microbial environment and clarifies how giant pandas are able to partially digest bamboo fibers despite a genome lacking enzymes that can degrade cellulose. It is becoming increasingly clear that giant pandas possess a suite of evolutionary adaptations for the highly specialized herbivory. Our findings offer a more in-depth look at the microbiome of a species that occupies an interesting place in the evolutionary tree and has an unusually narrow diet. Thus, the putative harboring of cellulose and hemicellulose-digesting microbes in the gut of the giant panda, along with a suite of other traits, including pseudothumbs; well-developed teeth, mandible and skull morphology, and chewing muscles; high-volume food ingestion (12.5 kg/day); brief digestion time; and high mucous levels in the digestive tract, likely have arisen as a result of adapting to a highly fibrous bamboo diet within the constraints imposed by the panda's innate carnivore-like digestive system.
Sample Collection.
Fecal samples from wild and captive giant pandas were collected immediately after defecation, snap-frozen in liquid N2, and shipped to the laboratory on dry ice. All samples were obtained from inside the feces, where there was no contact with soil. Six fecal samples from wild giant pandas were collected from the Qinling Mountains, and one sample was collected from the Xiangling Mountains by following GPS-collared animals or by directly observing individuals. Eight fecal samples from captive giant pandas (six animals were born in captivity, and two have been kept in captivity for 9 y) were collected from breeding centers between April and May 2008 and between April and May 2009.
DNA Extraction and PCR Amplification.
DNA was extracted from fecal samples using the Qiagen QIAamp DNA Stool Mini Kit according to the protocol for isolation of DNA for pathogen detection. DNA was eluted in a final volume of 250 μL using elution buffer and then stored at −20 °C. Tubes containing only the QIAamp DNA Stool Mini Kit extractions were included throughout the lysis and PCR steps to serve as negative controls. PCR amplification of 16S rRNA gene was performed as described previously (2). Five replicate PCRs were performed for each DNA sample. Each 25-μL reaction contained 60–100 ng of purified DNA, 12.5 μL of HotStarTaq Plus DNA Polymerase (Qiagen; 0.63 U DNA polymerase, 0.2 mM dNTPs, 2 mM Mg2+, 10 × PCR buffer), 0.5 μM Bact-8F (5′-AGAGTTTGATCCTGGCTCAG-3′), 0.5 μM Bact-1391R (5′- GACGGGCGGTGWGTRCA-3′), and 1 μg/μL BSA (Promega). Cycling conditions were as follows: 95 °C for 5 min, followed by 35 cycles of 95 °C for 40 s, 55 °C for 45 s, and 72 °C for 2 min, with a final extension period of 10 min at 72 °C. Replicate PCRs were pooled and gel-purified with the Wizard SV Gel and PCR Clean-Up System (Promega). Amplicons were cloned into PMD18-T vector (TaKaRa) and transformed into E. coli TOP10 (Tiangen). Approximately 192–900 colonies from each sample PCR product were chosen at random. Plasmid inserts were sequenced bidirectionally using BigDye Terminator (Applied Biosystems) and vector-specific primers (M13F: 5′-GTAAAACGACGGCCAG-3′; M13R: 5′-CAGGAAACAGCTATGAC-3′). Sequences were analyzed with DNASTAR (DNAStar) and trimmed to remove vector sequences. After trimming and adjusting for quality values, the average single sequence read length was ~700 nucleotides. Bidirectional sequence reads of clone inserts provided near–full-length 16S rRNA gene sequences of ~1,400 bp.
Chimera Checking.
Each sequence was edited manually in conjunction with its chromatogram and secondary structure information. Only unambiguous nucleotide positions were included in the analysis, and primer sequences were excluded. A multiple sequence alignment was generated with the NAST online tool (27), and chimeras were identified with Bellerophon version 3 (28), implemented at the Greengenes Web site (http://greengenes.lbl.gov) with the following (default) parameters: Sequences were compared with others within the same host species and with the Greengenes Core Set, identity to the core set was set to 97%, the match length to sequence threshold was set to 1,050 bp and 1,150 bp, respectively, the window size was set to 300, the count of similar sequences to search for each window was 7, the parent-to-fragment ratio was 90%, and the divergence ratio threshold was set at 1.1.
Determining OTU and Taxonomy Assignments.
Determining OTU for each sample. OTUs were defined as terminal nodes in phylogenetic analysis. Sequences remaining after the chimeras were checked for each individual were aligned using CLUSTAL in MEGA4 (29). OTU determination was performed according to DOTUR analysis (14). In DOTUR, sequences were grouped into OTU using distance-similarity matrices (by PHYLIP; http://evolution.genetics.washington.edu/phylip.html), such that the least similar pair within the OTU shared at least 97% identity.
Determining OTUs for all samples. To increase validity and decrease computing time, we determined all OTUs and shared OTUs in the sequences for all samples in two steps. First, we selected one wild and one captive sample sequence as the central library. We combined each sample sequence with this central library independently, and then determined the shared OTUs between the central libraries with each sample. Second, we combined unshared sequences for each sample into one new library, then used the foregoing method to determine the OTUs in this remaining large library. We chose a representative sequence for each OTU in each of the 97% identity groups at random, then used the BLAST method to find the closest GenBank neighbor for these representative sequences. Each representative sequence for each OTU was submitted to GenBank and granted an accession number. The result of EzTaxon (30), which was also used to determine the closest taxon for each OTU, served as the reference. Further, each representative sequence for each OTU with <95% identity to any GenBank sequence was defined as a previously undescribed OTU.
Phylogenetic Analysis.
Based on the foregoing taxon assignment results, we downloaded the representative of the 16S rRNA gene sequences of the bacteria from the National Center for Biotechnology Information (NCBI) Nucleotide Database, combining all OTUs to construct the neighbor-joining phylogenetic tree (1,000 bootstraps) using MEGA4 (29). We selected the representative sequence for each OTU at random. We used this tree to determine the phylogenetic position for each out, and used the representative sequence on each OTU to construct the neighbor-joining phylogenetic tree (1,000 bootstraps) with MEGA4 (29).
Relationship Between All OTUs and Known Clostridium Clusters.
We downloaded sequences from GenBank according to known Clostridium clusters (15). We then combined these sequences with all of the OTUs in this study and constructed the neighbor-joining phylogenetic tree (1,000 bootstraps) using MEGA4 (29). Based on phylogenetic information (in the same clade), we determined the relationships between all OTUs and known Clostridium clusters.
Estimation of Microbial Diversity.
We calculated the Good coverage estimate as [1 − (n/N)] × 100, where n is the number of singleton sequences and N is the total number of sequences for the sample analyzed (3). We used the rarefaction method in DOTUR (14) to explore the diversity of our clone libraries and found that the number of observed OTUs increased with additional sampling effort; however, the 99% coverage obtained over existing bacterial clone libraries indicated that only 1 new OTU would be expected for every 100 additional sequenced clones. Species richness was defined as the number of gut microbe species detected in a sample. DOTUR was used to calculate various diversity indices along with species richness. The classic Shannon–Weaver diversity indices (with 95% confidence interval) of each sample were calculated as described previously (14).
Metagenomics Analysis: DNA Extraction, DNA Library Construction, and Sequencing.
DNA was extracted from fecal samples of three individual wild giant pandas using the Qiagen QIAamp DNA Stool Mini Kit according to the protocol for isolation of DNA for pathogen detection. Sequencing and general data analyses were performed by BGI (Shenzhen, China). DNA library construction and sequencing followed BGI's previous work on human gut microbe metagenomic sequencing (7). A library with 350-bp clone insert size was constructed for our sample. Approximately 83 million high-quality reads, with read lengths of 90 bp, were generated for the samples; thus, the total data volume of high-quality reads was nearly 7.5 Gbp (SI Appendix, Table S3).
Public Data Use.
Public data used in the metagenomic analysis included the human gut contig set (version December 10, 2009), the NCBI sequenced bacteria genomes database (version January 13, 2010), the NCBI sequenced fungi genomes database (version March 17, 2010), the NCBI sequenced protozoa genomes database (version July 25, 2007), the Ribosomal Database Project (version July 14, 2010), and the integrated NCBI-NR database (version March 2011).
Illumina Genome Analyzer (GA) Short Reads de novo Assembly.
We compared the raw short reads with giant panda genome data to remove the host sequence. The clean reads thus obtained were assembled to obtain long contig sequences by the SOAPdenovo assembler (22), as used in human gut microbe metagenomic analyses. We tried different Kmer frequencies to obtain different assembly results, and used N50 lengths to access the best assembly result. The longest contig length and highest read utilization rate were obtained in Kmer55 for W1, Kmer31 for W2, and Kmer53 for W5. Thus, we used the contigs of Kmer55 for W1, Kmer31 for W2, and Kmer53 for W5 as the final assembly result (SI Appendix, Table S4).
Gene Prediction and Taxonomic Assignment.
We used the assembly contig sequences and applied MetaGene software, with only ORFs longer than 100 bp preserved. We translated the ORFs into protein sequences using NCBI Genetic Code 11. We carried out BLASTP (31) alignment to query the predicted protein sequences against the integrated NR protein database. For each predicted gene, hits with E-values >1 × 10−5 were filtered. Then a significant-matches set was retained to distinguish taxonomic groups, which were defined for hits with E-values <10 times the top hit E-value. Next, the LCA-based algorithm implemented in MEGAN (24) was introduced to determine the taxonomic level of each gene. The LCA-based algorithm assigns genes to taxa so that the taxonomic level of the assigned taxon reflects the level of conservation of the gene. For example, if the hits in the significant matches set belonged to one species, then the predicted gene was considered conserved within the species and assigned to the species. If the hits belonged to several species within one genus, then the predicted gene was assigned to the genus, which was considered the LCA of the predicted gene. If the hits belonged to more species from different genera and all of the genera belonged to one family (or higher taxonomic unit), then the predicted gene was assigned to the family, which was then considered the LCA. The remaining higher taxa can be determined in the same manner.
Gene Functional Classification.
We performed predicted gene functional classification by querying protein sequences of the genes against the eggNOG database (an integration of the COG and KOG databases) and the KEGG database using BLASTP with E-values <1 × 10−5. Genes were annotated as a function of the NOG or KEGG homologs with the lowest E-value. In the COG database, genes were classified into COG categories, whereas in KEGG, genes were assigned to KEGG pathways.
Glycoside Hydrolase (GH) Family Annotation and Analysis.
We performed database searches for GHs using HMMER HMM-based sequences search with Pfam hidden Markov models (Pfam v25.0 and HMMER v3.0) as described previously (20, 25). We used the Pfam_ls HMMs to find complete matches with the family by global alignment. All hits with E-values <10−4 were counted, and their sequences were analyzed further. GHs were named in accordance with the CAZy nomenclature scheme (32). For those GH families for which there is currently no Pfam HMM-based sequences, the representative sequences selected from the CAZy Web site were used in BLAST searches of the metagenomic data to identify these GH families using an E-value cutoff of 10−6 (20, 25).
We thank Drs. Chaodong Zhu and Hengling Cui for assistance with data analysis, Yonggang Nie for assistance in the field, the anonymous reviewers for their constructive comments, Prof. James Elser for the English polishing and giant panda breeding centers for sampling assistance. This study was supported by the National Basic Research Program of China (973 Program; 2007CB411600), the National Natural Science Foundation of China (30830020 and 31070331), and the Knowledge Innovation Program of the Chinese Academy of Sciences (KSCX2-EW-Z-4).
The authors declare no conflict of interest.
Data deposition: The sequences reported in this paper have been deposited in the GenBank database. For a list of accession numbers, see SI Appendix. The metagenomic data project ID is 5936 in Integrated Microbial Genomes with Microbiome Samples (IMG/M).
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/lookup/suppl/10.1073/pnas.1017956108/-/DCSupplemental.
Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
Full text links
Read article at publisher's site: https://doi.org/10.1073/pnas.1017956108
Read article for free, from open access legal sources, via Unpaywall:
https://www.pnas.org/content/pnas/108/43/17714.full.pdf
Citations & impact
Impact metrics
Citations of article over time
Alternative metrics
Smart citations by scite.ai
Explore citation contexts and check if this article has been
supported or disputed.
https://scite.ai/reports/10.1073/pnas.1017956108
Article citations
Gut microbial changes in a specialist blister beetle larvae and their nutritional metabolic characteristics.
Ecol Evol, 14(8):e70184, 22 Aug 2024
Cited by: 0 articles | PMID: 39184568 | PMCID: PMC11341433
Diversity, functions, and antibiotic resistance genes of bacteria and fungi are examined in the bamboo plant phyllosphere that serve as food for the giant pandas.
Int Microbiol, 21 Aug 2024
Cited by: 0 articles | PMID: 39168909
Multi-Omics Analysis Reveals Dietary Fiber's Impact on Growth, Slaughter Performance, and Gut Microbiome in Durco × Bamei Crossbred Pig.
Microorganisms, 12(8):1674, 14 Aug 2024
Cited by: 0 articles | PMID: 39203515 | PMCID: PMC11357262
Comparative analysis of energy homeostasis regulation at different altitudes in Hengduan Mountain of red-backed vole, Eothenomys miletus, during high-fat diet acclimation: examining gut microbial and physiological interactions.
Front Microbiol, 15:1434346, 10 Jul 2024
Cited by: 0 articles | PMID: 39050639 | PMCID: PMC11266106
Exploring Alashan Ground Squirrel (Spermophilus alashanicus) Diversity: Metagenomic and Transcriptomic Datasets from the Helan Mountains.
Sci Data, 11(1):517, 21 May 2024
Cited by: 0 articles | PMID: 38773139 | PMCID: PMC11109200
Go to all (220) article citations
Other citations
Data
Data behind the article
This data has been text mined from the article, or deposited into data resources.
BioStudies: supplemental material and supporting data
Pfam (Showing 34 of 34)
- (2 citations) Pfam - PF00959
- (1 citation) Pfam - PF00933
- (1 citation) Pfam - PF01229
- (1 citation) Pfam - PF03648
- (1 citation) Pfam - PF05592
- (1 citation) Pfam - PF07470
- (1 citation) Pfam - PF01270
- (1 citation) Pfam - PF03198
- (1 citation) Pfam - PF02065
- (1 citation) Pfam - PF01074
- (1 citation) Pfam - PF01055
- (1 citation) Pfam - PF02324
- (1 citation) Pfam - PF03632
- (1 citation) Pfam - PF02446
- (1 citation) Pfam - PF02449
- (1 citation) Pfam - PF00128
- (1 citation) Pfam - PF00722
- (1 citation) Pfam - PF01832
- (1 citation) Pfam - PF00703
- (1 citation) Pfam - PF00704
- (1 citation) Pfam - PF00728
- (1 citation) Pfam - PF01183
- (1 citation) Pfam - PF05089
- (1 citation) Pfam - PF00150
- (1 citation) Pfam - PF02056
- (1 citation) Pfam - PF00295
- (1 citation) Pfam - PF00251
- (1 citation) Pfam - PF01120
- (1 citation) Pfam - PF01464
- (1 citation) Pfam - PF00232
- (1 citation) Pfam - PF00331
- (1 citation) Pfam - PF01301
- (1 citation) Pfam - PF01204
- (1 citation) Pfam - PF03644
Show less
Similar Articles
To arrive at the top five similar articles we use a word-weighted algorithm to compare words from the Title and Abstract of each citation.